A comparison of pregnancy loss between transcervical and transabdominal chorionic villus sampling

To evaluate the comparative safety of transcervical and transabdominal chorionic villus sampling (CVS). From May 1988 to January 1992, CVS was performed by two operators at 9-12 weeks' gestation in 1048 singleton pregnancies. The sampling method for each patient, transabdominal or transcervical, was chosen primarily based upon placental location; the transabdominal route was used for anterior or fundal location and the transcervical route for posterior placentation. Perinatal outcome was assessed by post-procedure patient telephone contact, mid-gestation ultrasound evaluation, postpartum questionnaire completed by the referring obstetrician, and telephone interview with each patient after delivery. Complete follow-up was available in 1012 cases (97%). Excluding 39 elective abortions, 35 of 973 euploid pregnancies aborted spontaneously. The difference in fetal loss rate between transcervical and transabdominal CVS approached statistical significance (5.2 versus 2.9%; P = .058). Bleeding before CVS (P = .006) and multiple placental aspirations (P = .022) were associated with fetal loss for the entire study group. An interaction between uterine position and sampling method was also indicated; an increased loss rate was associated with transcervical CVS in the presence of uterine retroversion (P = .0017). Despite choosing the preferred CVS method for each patient, an increased loss rate may be associated with transcervical sampling in the presence of uterine retroversion.